Add 6

Dockerfile

@@ -1,42 +1,14 @@
-FROM nvidia/cuda:12.1.0-cudnn8-runtime-ubuntu22.04
+FROM python:3.10-slim
 
-# ======================
-# Environment
-# ======================
-ENV DEBIAN_FRONTEND=noninteractive
-ENV PYTHONUNBUFFERED=1
-ENV PIP_NO_CACHE_DIR=1
-
-# ======================
-# System dependencies
-# ======================
-RUN apt-get update && apt-get install -y \
-    python3 \
-    python3-pip \
-    git \
-    ffmpeg \
-    libgl1 \
-    ca-certificates \
-    && rm -rf /var/lib/apt/lists/*
-
-# python -> python3
-RUN ln -s /usr/bin/python3 /usr/bin/python
-
-# ======================
-# App setup
-# ======================
 WORKDIR /app
 
-# Copy & install python deps
+RUN apt-get update && apt-get install -y git && rm -rf /var/lib/apt/lists/*
+
 COPY requirements.txt .
-RUN pip install --upgrade pip && pip install -r requirements.txt
+RUN pip install --no-cache-dir -r requirements.txt
 
-# Copy app
 COPY app.py .
 
-# ======================
-# Runtime
-# ======================
 EXPOSE 7860
 
 CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

app.py

@@ -13,43 +13,63 @@ from diffusers import (
     EulerAncestralDiscreteScheduler
 )
 
+# =========================
+# Config
+# =========================
 MODEL_ID = "peter-sushko/RealEdit"
-
 FIXED_STEPS = 50
 FIXED_GUIDANCE_SCALE = 2.0
 
+# =========================
+# App
+# =========================
 app = FastAPI(title="RealEdit API")
 
 print("Loading RealEdit model...")
 
+# Detect device
+use_cuda = torch.cuda.is_available()
+device = "cuda" if use_cuda else "cpu"
+dtype = torch.float16 if use_cuda else torch.float32
+
+print(f"Using device: {device}, dtype: {dtype}")
+
+# Load pipeline
 pipe = StableDiffusionInstructPix2PixPipeline.from_pretrained(
     MODEL_ID,
-    torch_dtype=torch.float16,
+    torch_dtype=dtype,
     safety_checker=None
 )
 
 pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(
     pipe.scheduler.config
 )
-device = "cuda" if torch.cuda.is_available() else "cpu"
-pipe = pipe.to(device)
 
+pipe = pipe.to(device)
 
 print("Model loaded successfully!")
 
-
 # =========================
 # Core inference
 # =========================
 @torch.inference_mode()
 def run_inference(image, prompt):
-    with torch.autocast("cuda"):
+    if device == "cuda":
+        with torch.autocast("cuda"):
+            result = pipe(
+                prompt=prompt,
+                image=image,
+                num_inference_steps=FIXED_STEPS,
+                image_guidance_scale=FIXED_GUIDANCE_SCALE
+            ).images[0]
+    else:
         result = pipe(
             prompt=prompt,
             image=image,
             num_inference_steps=FIXED_STEPS,
             image_guidance_scale=FIXED_GUIDANCE_SCALE
         ).images[0]
+
     return result
 
 
@@ -122,4 +142,5 @@ gradio_ui = gr.Interface(
     )
 )
 
+# ⚠️ Mount UI at ROOT for Hugging Face Spaces
 app = gr.mount_gradio_app(app, gradio_ui, path="/")